Access MLLCL (and other variables) on USGS HPC Tallgrass

[putmanannie]

I've been accessing CONUS404 data via Tallgrass based on the hytest documentation, but have found that not all variables listed in the CONUS404 data release (322 variables, per wrfout_datadictionary.csv) or the publication (https://journals.ametsoc.org/view/journals/bams/104/8/BAMS-D-21-0326.1.xml) are available through this method.

For example, I would like to use the variable MLLCL (Mixed-layer lifting condensation level (LCL)) but it's not among the 153 data variables available when I access CONUS404 using the hytest specified method shown below. How might I find the any of the missing 170 data variables that are listed in the data release and publication but not available through the USGS HPC data access pathway?

`hytest_cat = intake.open_catalog("https://raw.githubusercontent.com/hytest-org/hytest/main/dataset_catalog/hytest_intake_catalog.yml")

cat = hytest_cat['conus404-catalog']
dataset = 'conus404-daily-osn' # have also checked conus404-hourly-osn, conus404-XXX-onprem doesn't appear to work
ds = cat[dataset].to_dask()
ds # to show view of dataset and explore variables`

[pnorton-usgs]

The online CONUS404 datasets only provide a subset of about 90 variables from the WRF model output. These variables were selected to constrain the overall size of the online dataset while providing variables commonly needed for studies within USGS. As you've found, MLLCL is one of the variables that currently is not included. There is no quick way to get additional variables from the CONUS404 model output. Either the additional variable(s) need to be added to the online dataset, or an enduser can pull the model output files themselves and pull out what they need; both cases require downloading the CONUS404 source model output files from Black Pearl (somewhere around 170TB of data). As additional use cases are identified for including other model variables we can update the online dataset depending on current priorities and resources.

@amsnyder what are your thoughts on this?

[amsnyder]

Hi @putmanannie - thanks for opening an issue! We can consider adding this data to the CONUS404 zarr dataset, but I will need to check on priority as we have a few other tasks that are high priority in our queue. Could you give me some context about what you are using this variable for (what kind of analysis/modeling are you doing?) and which projects you are using it for so that we can better determine where to prioritize this. Do you have any deadline for when you would need this data?

[putmanannie]

Hi @amsnyder --

The context is that I'm testing using a variety of Wx variables as local predictors of isotope ratios in precipitation in the UCRB to make a high temporal resolution isoscape as input for isotope enabled models to trace processes (SnowModel, maybe others, TBD). From past research, we know that condensation temperature can be an important predictor, and LCL has been shown to be a decent approximation of that for stable water isotopes. Hence my interest in testing the variable.

Fortunately, while I was poking around yesterday I found another method to get at the LCL. Metpy has a function (https://unidata.github.io/MetPy/latest/api/generated/metpy.calc.lcl.html) that does an approximation calculation using PSFC, T2 and TD2. That approximation should be sufficient for my needs for this particular project. So for the purposes of my project, I think it's not a priority to add it to the zarr dataset right now. I will let you know if I come across other variables that may be useful.

[aubreyd]

Hi @putmanannie - I believe this variable (MLLCL) is available in the RDA archive (labelled "MIXED-LAYER LCL").
https://rda.ucar.edu/datasets/d559000/

[putmanannie]

Hi @aubreyd -

Thanks for that info! I'll check it out.

best,
Annie